A follow-up to our previous note: "A Lack of Situational Awareness"
Last week we wrote about a market so absorbed in a $30 billion blow-up that it missed the more consequential move happening behind it: the US 30-year yield closing at 5.28%, its highest print in nineteen years, with Fed funds still parked at 3.50%-3.75%.
The result is a 150-basis-point gap between the policy rate and the 30-year, against a spread that sat close to zero in 2007. A gap of that size, opened without a given policy move behind it, has implications that run well past the AI trade and into central bank rate outlooks across the globe.
The US long bond has repriced the risk-free rate higher, in real terms, without assistance from the Fed. Given that real yields account for roughly 93% of the move in the 30-year since mid-June, this is more a term-premium and supply story, not a break-even inflation one. Thus, the market is not necessarily saying inflation will run away again; it is saying that the United States has to pay more to borrow because investors are less comfortable with the volume, composition, and future path of debt issuance.
The 5.2%–5.3% print on the 30‑year has been grabbing the headlines recently, however the more telling move over the past year has been in the 3‑ to 5‑year sector f the bond curve- exactly where Treasury refinancing, corporate rollovers, and AI‑related issuance are most concentrated. The 2007 comparison brings this into focus. Back then, the 30‑year traded roughly in line with Fed funds. Today, it’s about 150 basis points above the policy rate. That is not a curve that speaks to strong confidence in the central bank’s grip on the long end.
For central banks globally, this constitutes somewhat of a regime change. The long end of the US curve is beginning to drive financial conditions independently of the policy rate - the reverse of the framework's intended operation. If the US term premium continues to rise, it effectively exports tighter real financial conditions globally without a single hike being delivered. Central banks that expected to remain on hold, or to cut into a softening labour market, may find the curve delivering some of the tightening for them, or forcing a response purely to preserve credibility.
Steeper curves push up long-term funding costs and make investors pricklier about duration. The pressure shows up earliest in countries that rely heavily on external financing or have big rollover needs.
For emerging markets, and South Africa in particular, there is no opt-out. Higher US real yields compress the scope for EM central banks to cut, as each basis point of easing risks additional pressure on the currency and on capital flows. The SARB may retain a domestic case for lower rates, but the external constraint has tightened without any local action.
A steeper US curve, on top of our own fiscal worries, leaves us with a two-fold term-premium problem: global and domestic pressures hitting at the same time and feeding into one another. Duration in EM becomes a higher-conviction trade: greater confidence in the macro story is required to own the long end, and the compensation for being selective has increased accordingly.
Normalisation is no longer the conversation:
For long yields to normalise in a durable way, several conditions need to fall into place. We would need a credible narrowing of the US fiscal deficit. Investors would also need to regain confidence in the long‑term Treasury buyer base, especially as Japan trades more tactically and China continues to diversify its reserves. AI‑linked corporates would have to moderate supply in the same maturity buckets as Treasury refinancing. Finally, term premium would need to compress as investors grow more comfortable that this issuance is being absorbed.
None of that is in place.
Moreover, the rise in US long‑bond yields sits on a structural funding problem, not a standard cyclical adjustment. Tariffs have underdelivered as a relief valve: 2026 collections are running about a billion below earlier CBO projections, and between October 2025 and May 2026 the government raised roughly $189 billion from tariffs while spending about $742 billion on debt service. Spending cuts have not changed the trajectory either; even the headline savings are trivial against $39 trillion debt stock and a $7 trillion budget, and Congress has further rejected most structural proposals.
The foreign buyer base is also less reassuring. Japan, still the largest foreign holder at around $1.2 trillion, has at times been a forced seller when defending the yen, while China has cut its Treasury holdings to a 17‑year low and increased gold reserves, signalling a strategic shift away from US debt. On top of this, the AI capex boom is adding to supply in the same part of the curve: big US tech names are ramping issuance, with AI‑linked borrowing arriving alongside elevated sovereign supply in the belly of the curve.
Talk of “normalisation” has understandably faded. The problem now looks structural rather than cyclical. A Fed rate cut may ease conditions at the front end and buy some time for policymakers. It does not, however, fix the structural supply‑and‑demand imbalance in long‑dated debt, nor does it rebuild confidence in who will absorb that issuance over time.
First published on LinkedIn, 14 August 2026.